Multistage Ontology Research In Big Data Environment

With the development of technology and society in recent years, the global data is growing faster and faster and types of data are becoming more and more. This situation puts forward a challenge to the efficiently storing, quickly reading and retrieval of the data. As a kind of data representation, sharing and management methods, ontology is studied by more and more scholars. But there is a problem that once the ontology model is built, it is fixed.However, under the environment of Big Data, many new concepts are proposed every day,which should have been reflected in the old ontology model. At the same time, frequent updating is not only conducive to control the size of the ontology model, but also difficult to ensure the correctness.Faced with such a dilemma, this paper presents a new framework of multi-level ontology.In this framework, we can create a macro, global ontology model for big data level, and a micro-local ontology model for the underlying distribution data. The global ontology model is a relatively stable model, which reflects the general consensus and is built by experts in the area. It has higher authority and accuracy. And the local ontology model is a copy of the initial public body model. Then it will accept the huge amounts of data in the Web and expand new concepts to itself. And over a period of time, those concepts stabilized and becoming the new mainstream will be expanded to the global ontology model. Besides, we put forward a cooperative evolution strategies and the algorithm of conflict detection and resolution in the paper. By doing this, we can ensure that people can query the new concepts in time. Through such a approach, the framework ensure the stability and accuracy of the ontology model, and new concepts reflected in the ontology model timely.At the end of the paper, we design an experiment to verify the feasibility and superiority of the framework by ontology querying.